Aaditya Ramdas

A universal recipe for data-adaptive stopping rules unlocks anytime-valid guarantees for a broad spectrum of iterative algorithms, outperforming prior specialized methods.

Scaling a numeraire e-variable by an independent uniform multiplier guarantees a head-to-head win against any rival test with at least 50% probability, yet fundamentally degrades sequential power and expected log-growth.

Upper-bounding the number of distribution shifts in sequential data is provably impossible without parametric assumptions, but conformal inference can uniquely deliver valid, non-trivial lower bounds.